Patiënten die een colonresectie hebben ondergaan vanwege CRC hebben een verhoogd risico op een tweede metachrone tumor.

Het doel van endoscopische surveillance na een curatieve resectie van CRC is het detecteren en verwijderen van premaligne laesies en het detecteren van een lokaal recidief of metachroon CRC. Een juiste inzet van coloscopie surveillance na een in opzet curatieve chirurgische resectie vanwege CRC is wenselijk zodat de endoscopie capaciteit optimaal kan worden benut. Voor de surveillance van lokaal behandelde T1 CRCs verwijzen we naar de module T1 CRC van de landelijke richtlijn CRC.

Post-CRC endoscopic surveillance

The two main targets of endoscopic surveillance are early diagnosis of metachronous CRC and/or intraluminal recurrence of the index cancer.

Metachronous CRC

Epidemiology studies

In a Dutch study with 39 974 person-years of follow-up, the mean annual incidence of metachronous CRC was 0.3%. This corresponds to a cumulative incidence of 1.1% at 3 years, 2.0%at 6 years, and 3.1% at 10 years (standardized incidence ratio [SIR] 1.3, 95% confidence interval [CI] 1.1 – 1.5). This difference was observed particularly during the first 3 years following the initial CRC diagnosis (SIR 1.4, 95%CI 1.1 – 1.8)(Mulder, 2011). These data were confirmed by an absolute estimate of 1.8% for metachronous CRCs at 81 months. In a post hoc analysis of a large clinical study with 15 345 person-years of follow-up, a cumulative incidence of 1.5% (95%CI1.1%– 2.0%) was found for metachronous CRC at 5 years, corresponding to an SIR of 1.6 (95%CI 1.2 – 2.2) as compared to the general population (Green, 2002). In a French study including 61 879 person-years of follow-up, the incidence of metachronous CRC following the diagnosis of an initial CRC was higher than expected (SIR 1.5, 95%CI 1.3 – 1.7). It remained higher throughout the study period, significantly so between the first and the fifth years following diagnosis (SIR 1.9, 95%CI 1.6 – 2.3) (Bouvier, 2008). While direct comparisons are lacking, such excessive risk of metachronous cancer compared with the general population would appear to be of the same magnitude as that following resection of high-risk precancerous lesions (Atkin, 2017; Loberg, 2014).

The risk of metachronous CRC appears to be substantially higher in older studies or in those where preoperative colonoscopy was not systematically performed. For instance, in a clinical study from Korea, the risk of metachronous cancer in a population of 1049 subjects followed up for a mean of 41 months was approximately 0.2% per year (Lee, 2014). A similar estimate was observed in smaller Korean and UK studies (Moon, 2010; Patel, 2017). No metachronous cancer was detected in a French series where patients with metachronous cancer detected at perioperative colonoscopy were excluded (Barrier, 1998).

Endoscopic studies

The ESGE recently performed a systematic review of endoscopic studies in the setting of post-CRC surgery, in order to provide a reliable estimate of the adjusted incidence rate of metachronous CRC (Fuccio, 2018). The ESGE included 27 studies that enrolled 15 589 patients for a total of 15 803 index CRCs. The mean length of follow-up ranged from 18 to 108 months. The overall cumulative incidence of metachronous CRC was 2.2% (95%CI 1.8%– 2.9%).

Timing of metachronous CRC

Most studies indicate a higher risk for metachronous CRC in the initial years following surgery. A higher incidence in the first 2 – 3 years of follow-up has been reported in several epidemiological studies (Chen, 2013; Couch, 2013; Hyman, 2018; le Clercq, 2015; Mulder, 2012; Rulyak, 2007). This finding was also confirmed in our systematic review of endoscopic studies, which clearly showed that the risk of metachronous CRC was highest in the first 36 months following surgical resection and significantly decreased thereafter (Fuccio, 2018). Notably, about half of all metachronous CRCs discovered were diagnosed within 36 months and 70% within 60 months following surgical resection. This may be related to the quality of perioperative colonoscopy being undefined, as most of the series included in our systematic review commenced enrollment before 2000, with some as early as the 1970s.

In a large series of 3846 Chinese patients where a perioperative colonoscopy was systematically required to be performed, the majority of CRCs developed after 40 months from the index surgery (Lan, 2005). Similarly, other studies have shown low to no risk of early CRC in multiple cohorts with the systematic use of perioperative colonoscopy (Battersby 2014; Cone, 2013; Fuccio, 2015; Mafall, 2003; Rajaratnam, 2009; Sakamoto, 2013) however, a lead time bias cannot be excluded. For instance, a large study reported an inverse association between the time-interval of surveillance colonoscopy and the risk of metachronous CRC (Jayasekara, 2016).

Risk stratification

A few epidemiological and clinical studies have assessed whether the risk of metachronous CRC following surgery differs according to the main clinical or pathological features of the baseline cancer. Among the demographic variables, increasing age and sex were assessed and were not related to an increased risk of metachronous CRC (Bouvier, 2008; Jayasekara, 2016; Gervax, 2005; Lan, 2005; Marques-Antunes, 2017; Mulder, 2012). The site of the primary CRC being located proximally to the splenic flexure was associated with an increased risk of metachronous CRC in a population-based prospective study conducted among 5006 Swiss patients (Gervaz, 2005) and in a population-based study comprising 7863 subjects from the USA, Canada, Australia, and New Zealand (Jayasekara, 2016). However, a Dutch study including 10 283 patients (Mulder, 2012), a French study including 10 801 patients (Bouvier, 2008), and a recent Portuguese study including 535 patients (Marques-Antunes, 2017) have failed to confirm these findings. Conversely, a Spanish study reported an increased risk of metachronous adenomas and CRCs in patients with left-sided primary tumors (Borda, 2012). Absence of a high-quality baseline colonoscopy, defined as a complete examination with fair or good bowel cleansing, was linked to an increased risk of metachronous CRC in one small retrospective study (Marques-Antunes, 2017).

In conclusion, data on the risk factors for metachronous CRC are limited and often conflicting, therefore a recommendation for risk stratification for endoscopic surveillance cannot be made in this guideline.

Rate of early stage at diagnosis and cure of metachronous CRC

Studies consistently demonstrate that metachronous CRCs detected during colonoscopy surveillance are less advanced than index cancers and most are early stage lesions that are amenable to reoperation with curative intent (Barillari, 1996; Bouvier, 2008; Chen, 1994; Chen, 2000; Granqvist, 1992; Green, 2002; Hollington, 2011; Juhl, 1990; Kjeldsen, 1997; Marques-Antunes, 2017; Park, 2006; Schoemaker, 1998; Togashi, 2000). Kahi (2016) performed an analysis of pooled data from 31 studies, which showed that approximately two-thirds of metachronous CRCs are asymptomatic, detected at an early stage (TNM I-II, or Dukes A-B), and operated on with curative intent (Kahi, 2016). Radical surgery is possible more often for metachronous CRC than for local recurrence (67%– 86% vs. 7%– 22%, respectively) (Kjeldsen, 1997). Consequently, the survival from metachronous CRC appears not to be compromised relative to that observed in patients with a single CRC lesion (Fajobi, 1998; Papadopoulos, 2004).

Intraluminal recurrence

Timing of intraluminal recurrence

According to our meta-analysis, intraluminal CRC recurrences are more frequently detected endoscopically within the first 24 months postoperatively than in the more distant follow-up period (Fuccio, 2018). During the initial 2 years, the cumulative proportion of recurrent intraluminal cancers detected is 70.5%, with the highest rates being observed 6 – 12 months following CRC resection (recurrence in 1.7% of patients).

Furthermore, a separate study demonstrated that early postoperative colonoscopy is cost-effective. The number of 1- year colonoscopies needed to detect one CRC and to prevent one CRC-related death was 143 and 926, respectively. Moreover, the incremental cost of performing a 1-year colonoscopy (as compared to not performing it) is $ 40.313 per life-year gained (Hassan, 2009). Given these results, it appears that early surveillance colonoscopy (performed within 1 year following surgery) is economically justified.

Risk stratification

Tumor location

Although the incidence of malignant recurrence at the site of surgical anastomosis varies widely, reports consistently demonstrate a higher prevalence of rectal cancer recurrences (de Chaisemartin, 2009; Hojo, 1986; Jung, 2017; Kang, 2017; Kim, 2009; Matsuda, 2010; McDermott, 1985; Obrand, 1997; Patel, 2014) than colon cancer recurrences (Harris, 2002; Jung, 2017; Pihl, 1981; Obrand, 1997; Stulc, 1986). It has been estimated that more than 80% of all malignant recurrences occur in the rectum or distal colon (Kahi, 2016). This is further supported by the results of a recent meta-analysis that demonstrated a higher rate of endoscopically detected malignant relapses in patients with rectal cancer than with colon cancer (odds ratio [OR] 2.66, 95%CI 1.31 – 5.41). It also appears that the specific rectal cancer location (upper, middle, or lower rectum) is not related to the recurrence risk (Kim, 2009). For colon cancer, while one study failed to identify differences between left- and right-sided colonic tumors in terms of recurrence incidence (Jung, 2017), others observed that malignant recurrences are more common for primary tumors located in the left side of the colon (Bowne, 2005).

Other risk factors

Multiple studies have investigated factors other than tumor location on the risk of anastomotic or intraluminal CRC recurrence, yet they have largely produced inconclusive results . For example, tumor size larger than 5 – 6 cm (Hojo, 1986; Jung, 2017), stage T3 – T4 (Kim, 2009), and advanced nodal involvement (N2) (Jung, 2017) have been reported to influence tumor recurrence rates in rectal cancer, with other studies contradicting these associations (de Chaisemartin, 2009; Nash, 2010). With regards to histology, degree of differentiation (grade) appears not to influence the risk of malignant recurrence (Jung, 2017; Kim, 2009; Nash, 2010; Obrand, 1997). Nonetheless, it has been reported that the anastomotic and intraluminal recurrence risks are associated with mucinous differentiation of tumor cells (Hojo, 1986; Jung, 2017), lymphovascular invasion (LVI) (Hogan, 2015; Jung, 2017; Nash, 2010), and perineural invasion (PNI) (Jung, 2017), with the impact of LVI and PNI in rectal cancer being questioned.

It has been shown that, in patients undergoing total mesorectal excision, a clear distal margin of < 10 mm is adequate to prevent anastomotic recurrences in rectal cancer (de Chaisemartin, 20o9; Jung, 2017; Kang, 2017; Kim, 2009), with Nash et al. reporting otherwise when the margin is smaller than 8mm (Nash, 2010). Although involvement of the circumferential margin is an established risk factor for local relapses in rectal cancer (Adam, 1994; Nagtegaal, 2008), it appears not to influence the anastomotic recurrence rates specifically (de Chaisemartin, 2009; Kim, 2009).

Other factors studied in the context of their influence on anastomotic recurrence risk include intraoperative bowel lavage, use of staplers in rectal cancer surgery (Docherty, 1995; Matsuda, 2013; Terzi, 2006; Zhou, 2013), use of adjuvant treatment (de Chaisemartin, 2009; Jung, 2017; Kim, 2009; van den Brink, 2004), perioperative complications (Jung, 2017), and patient age, among others (de Chaisemartin, 2009; Jung, 2017; Kim, 2009; Obrand, 1997). With limited and conflicting evidence on the contribution of these factors to recurrence risk, no position can be taken on their use in stratifying patients with CRC to distinct colonoscopy surveillance strategies.

The main reason to perform surveillance colonoscopy at 1 year following surgical resection (or the perioperative colonoscopy) is the known early (within 12 – 24 months) intraluminal recurrence rate following CRC surgery. A similar pattern of early recurrence has also been demonstrated for metachronous CRC, which therefore further supports an early rather than delayed surveillance colonoscopy.

Intensive endoscopic surveillance

The benefit associated with intensive multimodality follow up of postoperative CRC patients remains controversial. Although most individual trials (Grossmann, 2004; Kjeldsen, 1997; Makela, 1995; Ohlsson, 1995; Rodriguez-Moranta, 1006; Rosati, 2016; Schoemaker, 1998) have failed to demonstrate a survival benefit with intensive follow-up, several older meta-analyses have reported an improvement in overall survival in intensively surveilled populations. Interestingly however there was no improvement in CRC-specific survival (Figueredo, 1003; Pita-Fernandez, 1025; Renehan, 2005; Tjandra, 2007).

In an attempt to bring clarity to this matter, two large follow-up trials were recently performed using computed tomography (CT) and carcinoembryonic antigen (CEA) testing in retrospective (Snyder, 2018) and prospective settings (examinations at 6, 12, 18, 24, and 36 months after surgery in the intensive group vs. at 12 and 36 months after surgery in the routine group) (Wille-Jørgensen, 2018). These studies failed to demonstrate any benefit of intensive follow-up, either in CRC-specific or overall survival, of patients following curative CRC surgery. The findings are in agreement with the results of a recent meta-analysis (Mokhles, 2016) and a Cochrane systematic review (Jeffery, 2016). While the aforementioned studies generally do not contain information on the specific impact of endoscopy on patient outcomes, it is reasonable to assume that the lack of benefit from intensive multimodality follow-up also relates to the absence of benefit from the intensive use of each of the individual examinations, including colonoscopy.

Of the limited number of studies that have separately analyzed colonoscopy follow-up, two meta-analyses have suggested that more frequent endoscopic examination does appear to

translate into additional survival benefit (Pita-Fernandez, 2015; Tjandra, 2007). One randomized controlled trial (RCT) evaluated the role of colonoscopy intensity specifically, performing examinations at 3-monthly intervals for 1 year, at 6-monthly intervals for the next 2 years, and once a year thereafter in an “intensive” follow-up group, and at 6 months, 30 months, and 60 months postoperatively in a “routine” follow-up group (Wang, 2009). While the authors of this study reported that more frequent examinations resulted in higher detection rates of asymptomatic recurrence, more curative surgeries, and improved survival in patients who had intra-luminal recurrences, they failed to observe improved overall survival in the intensive surveillance group.

Accordingly, in light of the evidence, the ESGE deem intensive surveillance in general and intensive endoscopic surveillance in particular not to be justified.

Second and subsequent colonoscopies

According to epidemiological studies, the long-term risk of metachronous CRC is not, or is only slightly, increased compared with the general population (Atkin, 2017; Bouvier, 2008; Green, 2002; Loberg, 2014; Mulder, 2012; Siegel, 2018). As

reported above, such risk is mainly concentrated in the early period following surgery, thereby supporting a missed lesion at the time of index colonoscopy being the most plausible reason. The sequence of two high quality colonoscopies – i.e. perioperative and 1 year after CRC surgery – appears to be an effective response to minimize such risk, at least in those patients without additional synchronous or metachronous lesions. The main reason for subsequent surveillance colonoscopies is therefore that CRC prevention is based on the detection of precancerous lesions. However, unlike in post-polypectomy surveillance, the issue of cancer prevention in CRC patients has rarely been addressed in the literature. Specifically, following CRC surgery, the risk of metachronous advanced adenomas, which may be considered to be a surrogate for CRC prevention, has been estimated in only a few studies and without direct comparison to any reference standard. For example, a Korean study estimated a 4.4% incidence of advanced adenomas at 41 months. Similarly, a ≤ 10% risk of advanced neoplasia has been estimated in smaller clinical studies (Cubiella, 2003; Moon, 2010; Patel, 2017), with a higher incidence described in patients with a history of left colonic resection (Fuccio, 2015; Yun, 2018).

These risks appear in general to be lower than, or equivalent to, those reported at post-polypectomy surveillance in patients with high-risk adenomas (Kim, 2018; Matsuda, 2009; Yoshida, 2016). As the ESGE guidelines on post-polypectomy surveillance suggest 3-year and 5-year intervals for the first and subsequent surveillance colonoscopies following resection of advanced adenomas (Hassan, 2013), these same recommendations may be applied following a 1-year surveillance colonoscopy.

Metachronous high risk neoplasia

Patients undergoing surgery for CRC may be diagnosed with an advanced adenoma that by itself requires additional surveillance. According to epidemiological studies, 4.5% of patients with CRC have a synchronous CRC (Mulder, 2011) and approximately 21% of patients have a synchronous adenoma (Bouvier, 2008). A recent endoscopy-based study found higher percentages of synchronous adenomas and advanced adenomas at 28% and 19%, respectively (Marques-Antunes, 2017). Interestingly, these percentages were even higher when the analysis was restricted to high quality baseline colonoscopy, reaching 38% for adenomas and 27% for advanced adenomas.

The role of synchronous CRC as a risk factor for metachronous CRC is controversial: some studies have reported synchronous CRC to be associated (Jayasekara, 2016; Mulder, 2012) with an increased risk of metachronous CRC, while others have not (Bouvier, 2008; Lan, 2005). Synchronous advanced adenomas have been reported to be associated with an increased probability of metachronous CRC (Borda, 2017; Marques-Antunes, 2017) however, this is based on small retrospective studies and therefore no firm conclusion can be drawn.

As a result, it is acceptable to propose a subsequent surveillance colonoscopy at 3 years following the second surveillance colonoscopy in those patients with additional advanced lesions at previous examinations.

Discontinuation of surveillance

In line with what has been reported for post-polypectomy surveillance, endoscopic surveillance is of limited, if any, bene- fit in patients ≥ 80 years of age owing to the significant risk of competing causes of mortality, such as cardiovascular and other malignant diseases (Yazdanyar, 2009). This also applies to younger patients with severe comorbidities that affect their expected life- span. Therefore, individualized recommendations should be based on general health status, comorbidity, and the findings at previous colonoscopies.

To answer the clinical question, the ESGE/ESDO guideline Endoscopic surveillance after surgical or endoscopic resection for colorectal cancer (Hassan, 2019) was used. The summary of literature below was adopted from this guideline.